Category Archives: software development

The new Google Chrome browser: a bad day for Firefox

The Firefox angle is what puzzles me about Google’s announcement that it is is launching a new open source browser. We should get to try it tomorrow; perhaps we’ll see that Google is successfully reinventing the browser. In particular, this is a part of what is sometimes dubbed the Google OS: the client for cloud applications running on Google’s servers:

Under the hood, we were able to build the foundation of a browser that runs today’s complex web applications much better.

Google is using some proven technology in the form of the Webkit rendering component (as used in Apple’s Safari). I imagine it can do a decent job. But why? From Google’s perspective, the browser market was shaping up nicely already. Microsoft’s IE has a still large but declining market share; Mozilla Firefox is growing, has a vibrant community, and relies on Google for the bulk of its income in return for making it  the default search engine – a deal which has just been extended for three years.

Now Google appears to be going head-to-head against Firefox. It won’t necessarily succeed; Firefox has lots of momentum and will be hard to shift. Equally, I doubt that Microsoft’s market share will decline significantly faster against a Google browser than it would anyway against Firefox.

The risk is that this will split the open source community.

As for Firefox, this can only be bad news. It has the embarrassment of relying on a major competitor for its income, and the knowledge that it is driving traffic to a company that will push users to switch to an alternative.

Maybe Google Chrome is so good that it will all make sense when we get to try it. For sure, it is an intriguing development for web applications and I’m looking forward to seeing how well Google can substantiate its claims that it is “much better” for the job of running them.

Delphi and C++ Builder 2009 are available to order

I’m choosing my words carefully, because although the CodeGear/Embarcadero site is now showing Delphi 2009 as the current version, if you click through to the order page it only offers a pre-order. Still, it must be done or thereabouts. US prices are as follows:

Delphi 2009: Pro $874.00;  Enterprise $1974.00; Architect $3474.00

C++ Builder 2009 costs the same; or you can get a bundle with both for a relatively small extra cost, eg. $1074.00 for Delphi and C++Builder Professional.

Curiously, an upgrade to Delphi 2009 Pro only costs $374.00 (57% discount), but an upgrade to Enterprise is $1274.00 (35% discount). I can’t make sense of this except on the basis that any product labelled “Enterprise” is presumed not price sensitive.

So what’s not in the Pro version? The Enterprise edition adds drivers for server databases in the dbExpress database framework, the DataSnap multi-tier application framework, and a full range of modeling diagrams. Architect bundles ER/Studio Developer Edition, Embarcadero’s database modeling tool, with support for a wide range of database servers.

In other words, the majority of Delphi’s features are in the Pro edition, which is really much the best value, though if you need DataSnap or client-server dbExpress then I guess you have no choice.

The big features here strike me as Unicode in the Visual Component Library; and the new language features, generics and anonymous methods. I’ve not yet looked at the product though, so watch this space.

Parts of EcmaScript 4 deemed unsound for the Web

This was the conclusion of an EcmaScript meeting in Oslo last month. Specifically, as Brendan Eich explains, three features – packages, namespaces and early binding – were considered too heavyweight unsuitable for a browser scripting language. Here is Eich’s “Executive summary”:

The committee has resolved in favor of these tasks and conclusions:

1. Focus work on ES3.1 with full collaboration of all parties, and target two interoperable implementations by early next year.

2. Collaborate on the next step beyond ES3.1, which will include syntactic extensions but which will be more modest than ES4 in both semantic and syntactic innovation.

3. Some ES4 proposals have been deemed unsound for the Web, and are off the table for good: packages, namespaces and early binding. This conclusion is key to Harmony.

4. Other goals and ideas from ES4 are being rephrased to keep consensus in the committee; these include a notion of classes based on existing ES3 concepts combined with proposed ES3.1 extensions.

This means that the evolution of JavaScript is now on a new path, focused for now on a more modest enhancement to the language called EcmaScript 3.1.

Given how loudly Eich protested about EcmaScript 3.1 last October, it is a surprising turn of events. Was Eich convinced by the arguments of Microsoft and Yahoo in support of a more lightweight JavaScript?

What this means is that JavaScript 2.0 won’t happen as previously envisaged. John Resig:

… you can forget a lot of what you learned about ECMAScript 4, previously. Many of the complicated concepts contained in the language have been tossed. Instead there is a considerable amount of effort going in to making sure that new features will be easily duplicable through other means.

Eich and Resig are keen to stress that JavaScript will still be a highly capable language. Still, the obvious conclusion is that this will be good for plug-ins which support more powerful languages: Adobe Flash, Microsoft Silverlight, Sun Java or Java/FX. Personally I’m disappointed.

It is also presenting Adobe with a tricky problem, as it implemented much of an earlier specification for EcmaScript 4 in ActionScript 3. Rather than being a standard language, as Adobe had planned, it looks like this will now be more of an Adobe language. I doubt this will have much practical impact on developers.

PS Brendan Eich has commented below.

Detailed article on Microsoft’s Midori published

The Software Development Times has an in-depth look at Midori by David Worthington, based on “internal Microsoft documents”:

SD Times has viewed internal Microsoft documents that outline Midori’s proposed design, which is Internet centric and predicated on the prevalence of connected systems.

Recommended if you are interested in what Microsoft is contemplating for a future OS. Note that there’s no official word on whether this is more than just another research project.

Technorati tags: , ,

Silverlight 2 threading issues, Quickstarts not working

I’ve been working on a Silverlight tutorial involving reading an RSS feed. Silverlight has a SyndicationFeed class which is meant to make this easy – as Microsoft’s Scott Barnes enthuses here.

It is handy, but I discovered that the Quickstart Barnes refers to does not work in Silverlight 2 Beta 2. The Quickstart section on Silverlight.net needs some work. Even if you get to this Quickstart via the link for Silverlight 2 Beta 2 examples on this page, it is soon apparent that it is actually for Silverlight 2 Beta 1. Click the Run It button and you’ll see that it asks for the older runtime.

The code doesn’t work in Beta 2 either; and as so often with thread-y stuff, it’s not immediately clear what’s going wrong. I got a blank page and the following message in the Debug output window in Visual Studio:

A first chance exception of type ‘System.UnauthorizedAccessException’ occurred in mscorlib.dll

In situations like this I recommend breaking on all CLR exceptions (Debug – Exceptions – check the Thrown box for Common Language Runtime exceptions in Visual Studio). Run again; and this time Visual Studio stops on the line which updates a Silverlight TextBlock:

feedcontent.Text += "* " + item.Title.Text + Environment.NewLine

with the message “Invalid cross-thread access”:

Rooting about a bit, I found this post from Karen Corby on changes in Silverlight 2 Beta 2:

HttpWebRequest’s delegates are called on a new non-UI thread.

  • Delegates were previously always called on the UI thread.
  • You must invoke back on to the UI thread if the data you’re retrieving will be consumed by a UI element.
  • For an example, see the updated networking post series (part one).

What this means is that you have two doses of asynchronous coding to think about if you use HttpWebRequest. First, the request itself; and second, in the code you write for the response handler if it needs to update the UI – which in most cases it will.

The example referenced by Corby shows a neat solution using a SynchronizationContext object, or you can use the Dispatcher class as explained by Wilco Bauwer here – he also draws attention to locking issues. See also Shawn Wildermuth’s post though note that CheckAccess is available despite what is said here.

This adds a significant dose of complexity to Silverlight coding. I’m not sure if any of this will change again in the final release.

I also noticed that VB coders are not well served by the Silverlight examples out there, which are overwhelmingly C#. Looks like this is the language of choice if you want an easy life.

What’s new in Delphi 2009

Today I viewed David Intersimone’s Live Webinar on what’s new in Delphi 2009, code-named Tiburon.

This is a Win32-only release. I think you will want it (if you use Delphi), if only for the new language-level features: generics, anonymous methods, and unicode strings. I grabbed a few screens from the presentation. Generics:

Unicode – here’s the TEncoding class:

and Unicode in action:

There are also some new components, such as a neat collapsible panel called TCategoryPanelGroup, TBalloonHints, and Office-2007 style ribbon controls.

The ribbon controls interested me because I am wary of Microsoft’s Office ribbon patent. CodeGear/Embarcadero seems to be wrapping Microsoft’s controls*, as used by the CMFCRibbon* classes, which as I understand it are not the actual controls used in Office 2007 but share their look and feel. You therefore have to agree to Microsoft’s license for the Fluent UI in order to use the controls.

There are also major changes to the DataSnap middleware but DavidI didn’t go into this much in the presentation.

During the Q&A at the end there were the inevitable questions: what about 64-bit (coming in a later version); what about Mac/Linux (nothing to announce); what about the dreadful online help (errrmmm we’re working on it); what about .NET (coming in a later version). Some of the language changes seem to be making ready for .NET 2.0 compatibility.

No announced release date; but the roadmap shows this as a 2008 release; and if it’s being webinared now that suggests it won’t be too long a wait.

Delphi is still absolutely my favourite Win32 development tool and this should be a strong release. At the same time, it is all rather old-school: win32, native code, fat client. You can do web applications in Delphi, and there is an updated “VCL for the Web” in this release, but why would you?

Nevertheless, if there are any Delphi developers still hanging on to Delphi 7 (the last version with the old IDE), perhaps these important language changes along with what is now a mature new-generation IDE will be sufficient to persuade them to migrate.

*Update: Although DavidI said that Delphi’s ribbon controls wrap Microsoft controls, Nick Hodges says here that this is not the case. He is probably right as I’m not sure what controls Delphi could wrap. If the MFC team could not use the actual Office controls, but had to create its own implementation, then I should think a third party would be in the same position. I wondered if the VCL was actually using the MFC code but I doubt that would be straightforward either. This may be a confusion caused by the licensing requirement.

Installing Visual Studio 2008 SP1

First, the patch removal tool, now officially called the “Visual Studio 2008 Service Pack Preparation Tool”.

Wait ages. Useless progress bar stuck at 50%. Wait some more.

Help – it’s asking for the Visual Studio DVD. Hey, at least it proves it is doing something. But where is it? Scrabble round office, eventually find .iso instead and mount it.

Wait ages. We are now over an hour into this install, and haven’t got past the preparation tool yet. Still useless progress bar stuck at 50%.

It’s done at last. Next, mount and run the SP1 iso:

http://www.microsoft.com/downloads/details.aspx?FamilyId=27673C47-B3B5-4C67-BD99-84E525B5CE61&displaylang=en

Wait again. This is when you need multiple computers in order to get on with your work.

Now Windows Update pops up to request a restart. I’m not sure if this is coincidence, or something triggered by the install. It doesn’t seem a good plan to restart when the SP1 installer is still chugging away, so I refuse.

Wait.

Done. Now the installer requests the inevitable Windows restart.

Still need to reinstall Silverlight beta 2:

http://www.microsoft.com/downloads/details.aspx?FamilyId=50A9EC01-267B-4521-B7D7-C0DBA8866434&displaylang=en

That one wasn’t too bad in comparison.

Overall: took longer than I’d like, but it worked.

Is Silverlight evil?

That’s really the question John Markoff is asking in this New York Times piece about the way the NBC Olympics streaming video is putting Microsoft’s browser plug-in on the map. Someone has even popped up to state that this is another go at technology monopoly:

“They’re still playing the same games,” said Michael R. Nelson, professor of Internet studies at Georgetown University. “It’s a way to lock up the content, and it’s not enabling as much innovation as we would like to see.”

It seems to me that Microsoft cannot win here. If the NBC servers were stuttering then it would prove that Silverlight isn’t up to the job and can’t compete. If Microsoft were not doing Silverlight, it would confirm that the company doesn’t grok the Internet and is still stuck in the world of Windows and Office.

As it is, the NBC Olympics streaming video seems to be working pretty well, judging by reports like this from CNET:

The picture quality is quite spectacular. The mist is so real it could not possibly have been photoshopped in there by the Chinese authorities to provide some extra menacing ambience. This makes YouTube seem like student video.

From a PR perspective, accusations of being evil is probably the best Microsoft can hope for.

But is Silverlight evil? I spoke to Dean Hachamovitch, the IE general manager, at the Mix conference in the Spring. My impression was that Microsoft has back-pedaled on improving browser-hosted JavaScript because it would rather see developers target Silverlight. There is some substance to the idea that Microsoft is promoting its own technology at the expense of open standards.

Still, without Silverlight would we all be using the OGG Theora format and the HTML 5 Video tag? In reality, the alternative is Adobe Flash; and the competition is if anything energising Adobe, which is not a bad thing.

Even if Silverlight achieves its aims, and becomes a widely used cross-platform runtime for web applications, I doubt we will see anything similar to the Windows/Office dominance of the nineties. Microsoft’s investment in Silverlight is conceding the point: that the future belongs to cross-platform clients, zero deployment, and both data and applications in the cloud.

According to Markoff’s piece, Adobe’s Kevin Lynch questions Microsoft’s cross-platform commitment, mentioning:

Microsoft’s decision to reserve certain features like 3-D effects and downloading for the company’s Windows Vista operating system.

I think this is a reference to the full WPF (Windows Presentation Foundation), which only runs on Vista and XP. I’m not sure what the “downloading” bit means; but Lynch is right, in that the relationship between WPF and Silverlight must be a subject of intense debate within Microsoft. Cross-platform .NET is a huge step to take, and there must be internal voices questioning why on earth the company is helping its customers to deploy applications that are not tied to the Windows client.

Still, I think that bridge has now been crossed. Nobody wants to build Internet applications that only work on Windows; it isn’t an option. Therefore I doubt that Microsoft will hobble Silverlight in order to promote WPF; doing so would only help competitors. I may be proved wrong; it’s smart of Lynch to draw attention to Microsoft’s conflicted interests.

Silverlight is great work, and only a little bit evil.

Microsoft Silverlight: 10 reasons to love it, 10 reasons to hate it

A year or so a go I wrote a post called Adobe AIR: 10 reasons to love it, 10 reasons to hate it. Here’s the same kind of list for Microsoft’s Silverlight, based on the forthcoming Silverlight 2.0 rather than the current version. The items are not in any kind of order; they also reflect my interest in application development rather than design. It is not a definitive list, so there are many more points you could make – by all means comment – and it will be interesting to have another look a year from now when the real thing has been out for a while.

This Silverlight developer chart is available in full on Brad Abrams’ blog here, or in Joe Stegman’s Deep Zoom version here.

The pros…

1. The Silverlight plug-in means developers can target a single, consistent runtime for browser-based applications, rather than dealing with the complexity of multiple browsers in different versions. You also get video and multimedia effects that are hard or impossible with pure HTML and JavaScript; though Adobe’s Flash has the same advantages.

2. Execute .NET code without deploying the .NET runtime. Of course, the Silverlight plug-in does include a cut-down .NET runtime, but instead of dealing with a large download and the complexities of the Windows installer, the user has a small download of about 4MB, all handled within the browser. In my experience so far, installation is smooth and easy.

3. Performance is promising. Silverlight comes out well in this prime number calculator, thanks no doubt to JIT compilation to native code, though it may not compare so well for rendering graphics.

4. Support for Mono (Moonlight) means there will be an official open source implementation of Silverlight, mitigating the proprietary aspect.

5. Silverlight interprets XAML directly, whereas Adobe’s XML GUI language, MXML, gets converted to SWF at compile time. In fact, XAML pages are included as resources in the compiled .XAP binary used for deploying Silverlight applications. A .XAP file is just a ZIP with a different extension. This also means that search engines can potentially index text within a Silverlight application, just as they can with Flash.

6. Third-party component vendors are already well on with Silverlight add-ons. For example, Infragistics, ComponentOne and DevExpress.

7. Take your .NET code cross-platform. With Macs popping up everywhere, the ability to migrate VB or C# code to a cross-platform, browser-based Silverlight client will be increasingly useful. Clearly this only applies to existing .NET developers: I guess this is the main market for Silverlight, but it is a large one. The same applies to the next point:

8. Uses Visual Studio. Microsoft’s IDE is a mature and well-liked development environment; and since it is also the tool for ASP.NET, you can use it for server-side code as well as for the Silverlight client. For those who don’t get on with Visual Studio, the Silverlight SDK also supports command-line compilation.

9. Choose your language. Support for multiple languages has been part of .NET since its beginning, and having the .NET runtime in Silverlight 2.0 means you can code your client-side logic in C#, Visual Basic, or thanks to the DLR (Dynamic Language Runtime) Iron Ruby or Iron Python.

10. Isolated storage gives Silverlight applications local file access, but only in a protected location specific to the application, providing a relatively secure way to get this benefit.

The cons…

1. If Apple won’t even allow Flash on the iPhone, what chance is there for Silverlight?

2. Silverlight is late to the game. Flash is mature, well trusted and ubiquitous; Silverlight only comes out of beta in the Autumn (we hope) in the version we care about – the one that includes the .NET runtime – and will still lack support on mobile devices, even Windows Mobile, though this is promised at some unspecified later date.

3. The design tools are Expression Blend and Expression Design – but who uses them? The design world uses Adobe PhotoShop.

4. While having solution compatibility between Expression Blend and Visual Studio sounds good, it’s actually a hassle having to use two separate tools, especially when there are niggling incompatibilities, as in the current beta.

5. No support for the popular H.264 video codec. Instead hi-def video for Silverlight must be in VC-1, which is less common.

6. It’s another effort to promote proprietary technology rather than open standards.

7. Yes Linux will be supported via Moonlight, but when? It seems likely that the Linux implementation will always lag behind the Windows and Mac releases.

8. Silverlight supports SOAP web services, or REST provided you don’t use PUT or DELETE, but doesn’t have an optimized binary protocol like Adobe’s AMF (ActionScript Message Format), which likely means slower performance in some scenarios.

9. Silverlight is a browser-only solution, whereas Flash can be deployed for the desktop using AIR (Adobe Integrated Runtime). Having said that, yes I have seen this.

10. You have to develop on Windows. This is particularly a problem for the Expression design tools, since designers have a disproportionately high number of Macs.

SQL Server 2008 is done

Microsoft has announced that SQL Server 2008 is released to manufacturing – ie. the bits are done, even if you can’t buy it yet. MSDN subscribers can download it now.

This is the product that was “launched” back in February; it’s been a long delay but I get the impression that the SQL team likes to wait until its release really is ready.

SQL Server 2008 is more like a suite of products than a single product now. It has a large range of editions from Compact to Enterprise, and product areas like Analysis Services and Reporting Services are distinct from the core engine.

The pieces that interest me most are the spatial data types, sparse columns, FILESTREAM data type, and the various object-relational layers including LINQ, Entity Framework, ADO.NET Data Services, and the ongoing work with SQL Server Data Services (which is far from done yet).

DBAs will likely have a very different view of what is important, as will Business Intelligence specialists.

SQL Server has prospered by being cheaper than than the likes of Oracle and DB2, and by integrating smoothly with Windows and Active Directory. I wonder if it will feel pressure from even more cost-effective open source offerings like MySQL, as they become more Enterprise-ready?